Digital Therapeutics: Implications for Precision Treatment ...

Sabine Wilhelm, Ph.D.Massachusetts General Hospital,

Harvard Medical School

Digital Therapeutics:

Implications for Precision Treatment of

Psychiatric Disorders

DISCLOSURES

• Guilford Publications, New

Harbinger Publications, Oxford

University Press, Springer, Elsevier

Publications

• Brattleboro Retreat, International

Obsessive Compulsive Disorder

Foundation, Tourette Association of

America, Association for Behavioral

and Cognitive Therapies

• National Institute of Mental Health

• Koa Health, Inc.

• Noom, Inc.

• One Mind (PsyberGuide)

ROAD MAP

Global Mental Health

Crisis

Technology as a

Solution

Bias and Diversity

Considerations

Chatbots, Social Media

Platforms, Virtual Reality

Comprehensive Mental

Health Platform

Internet & Smartphone

Solutions

OF THOSE WITH 12-MONTH PSYCHIATRIC DIAGNOSES

60%

40%

No Care Care

Wang et al., 2005

38.9% received at

least “minimally

adequate” care

BARRIERS TO MENTAL HEALTHCARE

PATIENT-LEVEL BARRIERS

Logistical Issues

Financial Barriers

Shame & Stigma

1. Kass et al., 2017; 2. Crow et al., 2013

PROVIDER SHORTAGE

Lin, Stamm, Christdis, 2016

ROAD MAP

Global Mental Health

Crisis

Technology as a

Solution

Bias and Diversity

Considerations

Chatbots, Social Media

Platforms, Virtual Reality

Comprehensive Mental

Health Platform

Internet & Smartphone

Solutions

ROAD MAP

Global Mental Health

Crisis

Technology as a

Solution

Bias and Diversity

Considerations

Chatbots, Social Media

Platforms, Virtual Reality

Comprehensive Mental

Health Platform

Internet &

Smartphone Solutions

INTERNET CBT

GUIDED I-CBT IS EQUIVALENT TO IN-PERSON CBT

G=0.05 (95% CI: -0.09 to 0.20)

Carlbring et al., 2018

• Major Depressive Disorder (g = 0.67)

• Panic Disorder (g = 1.31)

• Seasonal Affective Disorder (g = 0.92)

• Generalized Anxiety Disorder (g = 0.70)

• Post-Traumatic Stress Disorder (g = 0.71)

I-CBT VS. CONTROL

1. Andrews et al., 2018; 2. Sijbrandij et al., 2016

Your smartphone is always-on, always-with-you, and knows where you are

52%

Mexico

55%

Philippines

60%

Brazil

60%

S. Africa

Pew Research Center, 2019

=76% =81%=78%

SMARTPHONE PENETRATION

PHONE-BASED TREATMENTS ADDRESS KEY BARRIERS

Many in Need

Long Wait Times

Logistical

Barriers

Shame & Stigma

Variable Quality

High Cost

THE MENTAL HEALTH APP MARKET

Marshall et al,, 2019

Marshall et al., 2019

There are over

10,000 apps

dedicated to

mental health

Symptom Monitoring

T2 Mood

Tracker

Single Skill

Daily Feats

With a Therapist

Recovery

Record

Full-Fledged Treatment

Perspectives

0

0.2

0.4

0.6

0.8

1

vs Inactive Controls vs Active Controls

Anxiety Depression

g = .56

[0.38, 0.74]g = .45

[0.30, 0.61]

g = .19

[0.07, 0.31]

g = .22

[0.10, 0.33]

Firth et al., 2017a-b

EFFICACY OF APPS

APPS ARE USEFUL FOR IMPROVING:

• Depression (g = .28)

• Generalized Anxiety (g = .30)

• Social Anxiety ( g = .58 )

• General Psychiatric Distress (g = .40)

• Stress Level (g = .35)

• Positive Affect (g = .44)

• Quality of Life (g = .35)

WHAT DO THEY WORK FOR?

Professional guidance significantly

increases effect sizes

0.48

0.53

0.57

0.52

0.230.21

0.24

0.29

0

0.1

0.2

0.3

0.4

0.5

0.6

Depression Anxiety Stress Level Quality of

Life

Yes No

Lindardon et al., 2019

p < 0.001

LEARNING CURVE: EVIDENCE

RESEARCH TRIALS PER DEPRESSION APP

Martínez-Pérez, de la Torre-Díez, & López-Coronado, 2013

LEARNING CURVE: ENGAGEMENT

CURRENT TRENDS IN ENGAGEMENT

0

20

40

60

80

100

120

Downloaded Next Day Next Week Next Year

Thousa

nds

iOS Android

Owen et al., 2015

COMMON REASONS FOR APP DISCONTINUATION

POOR USABILITY• Unhelpful in emergency

• Difficult/unenjoyable to use

• Takes too long to enter data

• Discovered hidden costs

PRIVACY CONCERNS• Concerns about data privacy (59%)

• Only 8% will share data with a tech company

• Lack of transparent privacy policies

LACK OF TRUST• Questions of effectiveness or

usefulness (71%)

• Bold claims create skepticism

• Too many apps/lack of regulation

1. Torous et al., 2018; 2. Lipschitz et al., 2019; 3. Torous et al., 2019

23%

29%

18%

12%

27%

25%

38%

34%

0%

5%

10%

15%

20%

25%

30%

35%

40%

CBT Mindfulness Mood Monitor Real-person

feedback

Yes No

Treatment Element and Dropout Rate

p = 0.037 p = 0.003

AI

Therapist Assisted

Peer-SupportAI Therapists

SimSensei

Chatbots

STAKEHOLDER INVOLVEMENT FROM THE START

1. Schueller et al., 2013; 2. Sucula et al., 2017

67%

33%

No Input

Input from Healthcare ProfessionalsPatients

Clinicians

Designers

Engineers

3rd

Party

Payers

PERSONALIZATION:JUST IN TIME INTERVENTIONS

SMARTPHONE SENSOR-BASED TECHNOLOGY

WiFi: Location

Heart Rate

Microphone:

Vocal Tone

GPS:

Location

Bluetooth:

Social Proximity

Camera:

Facial Expression

Outdoor

Light

Accelerometer:

Movement

Torous, 2017

Screen

on/off:

Sleep

JUST IN TIME INTERVENTIONS 1

• Great potential

• Active (short questionnaires) and

passive (sensor) data can be used to

prompt personalized real time

interventions

• Currently only 1% of marketplace apps

support sensor use

• This will change soon

Torous et al, 2021

LEARNING CURVE: SECURITY

RISK TO USERS’ PRIVACY ACROSS 43 HEALTH APPS

40%32%

28%

High risk Medium risk Low riskNjie, 2013

ROAD MAP

Global Mental Health

Crisis

Technology as a

Solution

Bias and Diversity

Considerations

Chatbots, Social Media

Platforms, Virtual Reality

Comprehensive Mental

Health Platform

Internet &

Smartphone Solutions

DIVERSITY CONSIDERATIONS

African American and Lantix use

smartphones to access health

information more frequently than

Whites

Digital health interventions could

alleviate mental health disparities

Underserved populations suffer the poorest health outcomes

1. Anderson-Lewis et al., 2018; 2. Pew Research Center, 2019:

3. Schueller, Hunter, & Figueroa, 2019 4. Fonagy & Luyten, 2021

82%80% 79%

82%

58% 57%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

White African American Hispanic/Latino

Smartphone Ownership Desktop/Laptop

African -American: 75% own a

smartphone, and 23% own a

cellphone

Latinx, 77% own a smartphone,

and 20% own a cellphone

TECHNOLOGY DEMOGRAPHICS

Home Internet Access

1. Schueller, Hunter, & Figueroa, 2019; 2. Pew Research Center, 2019

LGBTQ: Higher rates of

seeking health information and

resources online

Rural Populations: 65% own a

smartphone, 61% have home

internet access

Homeless Individuals: 58%

own a smartphone, 86% have an

Android

Messages must be designed to

meet the literacy, language,

cultural and motivational

needs of the population

63%

79%

50%

66%

49%

61%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

2010 2013 2016 2019

White African American Hispanic/Latino

ROAD MAP

Global Mental Health

Crisis

Technology as a

Solution

Bias and Diversity

Considerations

Chatbots, Social Media

Platforms, Virtual Reality

Comprehensive Mental

Health Platform

Internet & Smartphone

Solutions

INNOVATIVE DIGITAL INTERVENTIONS1

• Chatbots

• Automated conversational interfaces, can detect and respond to immediate needs

• Some individuals feel more comfortable conversing with a chatbot

• Widely used in many commercially available apps

• Social media Platforms

• Patterns of social media use can be used to detect individual worsening of symptoms (i.e., Schizophrenia relapse)

• Platforms use personalized therapy combined with social connections (i.e., Moderated Online Social Therapy platform)

• Virtual Reality

• Can create and control personalized, real-world exposures in a safe, convenient environment

• Can be used for mindfulness and relaxation1Torous et al, 2021

AREAS OF IMPROVEMENT1

• Chatbots

• Limited in recognizing serious mental health concerns and providing appropriate responses (i.e., suicidal ideation, domestic violence)

• Privacy concerns

• Social media Platforms

• Stigma and disinformation are widespread

• Ethical concerns when using social media for treatment

• Virtual Reality

• Mostly inaccessible and unfamiliar to the general public

1Torous et al, 2021

ROAD MAP

Global Mental Health

Crisis

Technology as a

Solution

Bias and Diversity

Considerations

Chatbots, Social Media

Platforms, Virtual Reality

Comprehensive Mental

Health Platform

Internet & Smartphone

Solutions

Peer

support

Virtual Reality

In-app coach

Live therapist

ASSESSMENTLevel of Care

PERSONALIZED

TREATMENT/PREVENTION

PACKAGE

Machine Learning/AI

SUMMARY AND CONCLUSIONS

• Mental Health Crisis

• Most individuals with mental health issues are not getting care

• Technology

• Offers scalable approaches to assessment and

personalized treatment

• Will continue to evolve rapidly

• Need to minimize risks and maximize benefits

for our patients.

ACKNOWLEDGEMENTS

Hilary Weingarden Nick Jacobson Jennifer Greenberg

Anna Schwartzberg Julia Carrellas Rebecca Berger-Gutierrez